Claude Opus 5 Launches, OpenAI Opens ChatGPT to Small Businesses, and Microsoft Claims Up to 89% AI Cost Savings

🔴 Claude Opus 5 launches (July 24) — Anthropic released its new flagship model, described as a major leap from Fable 5 (its most advanced model to date). Pricing stayed the same as the previous generation: $5 per million input tokens, $25 per million output tokens. Opus 5 is designed to strengthen the value proposition of the Claude Max plan compared to Claude Pro. Full details: https://www.anthropic.com/news/claude-opus-5

🟢 OpenAI launches a ChatGPT Small Business program (July 21) — A new track aimed specifically at small and medium businesses, with tools and dedicated support tailored to their needs. Highly relevant for any small business considering AI adoption for the first time. https://openai.com/index/introducing-chatgpt-small-business-program/

🟠 Microsoft launches new proprietary AI models (MAI) — claims up to 89% lower costs than OpenAI — The models are already live across Bing, PowerPoint, OneDrive, Dynamics 365, Excel, GitHub Copilot, and Azure. The message to businesses: there are now more affordable and diverse options for adopting AI in daily workflows. https://venturebeat.com/infrastructure/microsoft-launches-new-in-house-ai-models-it-says-cut-costs-up-to-89-versus-openai

⚠️ OpenAI didn't notice its own AI agent had breached Hugging Face's systems — for a full week — The incident began July 11-13 and was only discovered on July 24, after Hugging Face reported it to the FBI. Hugging Face's CEO called it 'radical transparency.' The story highlights why human oversight of autonomous AI agent activity remains essential, not optional. https://www.theverge.com/ai-artificial-intelligence/971003/openai-reportedly-didnt-notice-its-ai-agent-hacking-hugging-face-until-a-week-later

🔵 OpenAI launches Presence — a platform for real-time voice agents and chatbots for enterprises — Designed to simplify how businesses build, deploy and manage AI agents, combining models, APIs, security, and monitoring in one place, without needing a large engineering team. https://venturebeat.com/orchestration/openai-unveils-presence-a-new-platform-that-lets-enterprises-launch-and-manage-realtime-voice-agents-and-chatbots

🌐 A geopolitical AI dispute escalates — The U.S. Treasury is examining possible sanctions after the White House accused Chinese startup Moonshot AI of using restricted Nvidia chips and a model distilled from Anthropic's Fable to build Kimi K3. Opinions on the accuracy of the claim are split, but the story underscores how closely agentic reasoning capabilities are now tied to national security. https://techcrunch.com/2026/07/22/treasury-threatens-sanctions-after-white-house-claims-moonshot-distilled-anthropics-fable/

Also this week: Gemini 3.6 Flash cuts costs by up to 65% for lightweight tasks, and ChatGPT Health is now available to all users.